Static Stability: The Secret to High Availability

Two strategies you can use

Mar 18, 2025

When it comes to building highly available systems, every second of downtime can translate into financial losses, frustrated users, and damage to reputation. Large companies like Amazon, Google, and Netflix invest heavily in ensuring their systems remain operational even in the face of failures.

But what’s the secret behind their approach?

It’s a concept called Static Stability: a proactive strategy that ensures systems remain functional even when infrastructure components fail. While costly, it’s a must-have for mission-critical applications.

What’s Static Stability?

Imagine your system is running in three availability zones (AZs) in a cloud provider like AWS. A typical reactive approach would detect an outage in one AZ and then spin up new instances in a healthy AZ to maintain system functionality.

Static stability, however, proactively provisions extra resources from the beginning so that even if an AZ goes down, the system continues to operate at full capacity.

This approach is commonly used in mission-critical applications, where downtime is simply not an option.

Implementing Static Stability

1 - Active-Active High Availability

A classic way to implement static stability is through an Active-Active HA setup. This approach is commonly used in public-facing applications, such as web services, APIs, and mobile apps.

How It Works

You deploy your service across multiple availability zones (AZs).
A load balancer distributes traffic evenly across all instances.
Instead of provisioning only what is needed (e.g., 2 instances), you over-provision by at least 50% (e.g., 3 instances).
If one AZ goes down, the remaining instances can handle 100% of the load.

Example Scenario

Let’s say your service typically requires 2 application instances to handle incoming traffic. To achieve static stability, you deploy 3 instances across 3 AZs:

✅ 1 instance per AZ → Ensures no single point of failure.
✅ Each instance can handle 50% of total traffic → Even if one instance fails, the system remains operational.

2 - Active-Passive High Availability

Not all systems can run in Active-Active mode. Databases and stateful applications require Active-Passive HA, where only one instance is actively handling requests at a time.

How It Works

A Primary (Active) instance handles all reads and writes.
A Standby (Passive) instance sits in another AZ.
If the Primary fails, the Standby is promoted as the new Primary.
This ensures minimal downtime while maintaining data consistency.

Example Scenario

Imagine a critical financial database. A typical failover setup might look like this:

Primary Database (Active) – AZ-1
Standby Database (Passive) – AZ-2
Failover Mechanism (Automatic/Manual Switch)

The Trade-Off: Is Static Stability Worth It?

Many developers hesitate to implement static stability because of higher resource costs. However, the cost of downtime can often outweigh the cost of extra infrastructure.

When Should You Use Static Stability?

✅ Mission-Critical Systems – e.g., banking, healthcare, real-time monitoring systems.
✅ Highly Scalable Applications – e.g., global SaaS platforms, social media networks.
✅ Applications with SLAs (Service Level Agreements) – e.g., Enterprise-grade cloud services like AWS EC2, S3, RDS.

When Is It NOT Needed?

❌ Small applications with low uptime requirements.
❌ Development or testing environments where cost savings matter more.
❌ Systems that can tolerate downtime (e.g., internal dashboards).

👉 So - have you leveraged Static Stability for your applications?

System Design Codex

Discussion about this post